Influence of landmark and surgical variability on virtual assessment of total knee arthroplasty

Given advances in recent years in imaging modalities and computational hardware/software, virtual analyses are increasingly valuable and practical for evaluating total knee arthroplasty (TKA). However, the influence of variabilities at each step in computational analyses on predictions of TKA performance for a population has not yet been thoroughly investigated, nor the relationship between these variabilities and expected variations in surgical practice. Understanding these influences is nevertheless essential for ensuring the clinical relevance of theoretical predictions. Here, a morphological analysis of proximal tibial resections within TKA is proposed and investigated. The goals of this analysis are to quantify the influence of variability in landmark detection on resection parameters and to evaluate this sensitivity relative to expected clinical variability in TKA resections. Results here are directly applicable to population-level computational analyses of morphological and functional TKA performance.